Scheduling cron jobs with pg_cron and pg_net

Running background jobs directly in Postgres is surprisingly effective for many workloads. pg_cron schedules SQL at the database level, and pg_net can call HTTP endpoints from inside Postgres. Together they let you automate tasks without a separate worker.

Install the extensions

CREATE EXTENSION IF NOT EXISTS pg_cron;
CREATE EXTENSION IF NOT EXISTS pg_net;

Make sure pg_cron is configured with the right database in postgresql.conf.

Example: nightly cleanup

SELECT cron.schedule(
  'nightly_cleanup',
  '0 2 * * *',
  $$DELETE FROM audit_logs WHERE created_at < now() - interval '90 days'$$
);

Example: ping a webhook

SELECT net.http_post(
  url := 'https://example.com/webhooks/daily-rollup',
  headers := '{"Authorization": "Bearer token"}'::jsonb,
  body := '{"source":"pg_cron"}'::jsonb
);

Observability tips

• Store cron.job_run_details in a dashboard table.
• Add a short retention policy for old runs.
• Emit alerts for consecutive failures.

When to use this pattern

If your task is data-local and short-lived, pg_cron is a clean fit. For long-running or CPU-heavy jobs, keep a separate worker process.